The present invention relates to a method and system for polarization mode dispersion compensation of optical signals. In particular, the present invention relates to a method for polarization mode dispersion compensation using at least two linearly chirped Bragg gratings to selectively tune the reflection points of two polarization resolved signals, creating a variable polarization dependent delay.
Present day telecommunication systems require that optical signals be conveyed over very long distances. In an optical communications signal, data are sent in a series of optical pulses. Real signal pulses are composed of a distribution of wavelengths and polarizations, each of which travels at its own characteristic velocity. This variation in velocity leads to pulse spreading and thus signal degradation. Degradation due to the wavelength dependence of the velocity is known as chromatic dispersion, while degradation due to the polarization dependence is known as polarization mode dispersion.
Mathematically, the speed of light v in a waveguide is given by                     v        =                  c          n                                    (        1        )            where c is the velocity of light in free space and n is the effective index of refraction in the waveguide. Normally, the effective index, n, of the optical mode is dependent upon the wavelength. Thus components of light having different wavelengths will travel at different speeds. In addition to being dependent upon wavelength, the effective index in a waveguide may also be dependent upon the polarization of the optical signal. Even in “single-mode” fiber, two orthogonal polarizations are supported, and, in the presence of birefringence, the polarizations travel at different speeds. Birefringence in the fiber may arise from a variety of sources including both manufacturing variations and time-dependent environmental factors. The speed difference results in a polarization-dependent travel time or “differential group delay” (DGD) between the 2 different polarization modes within the birefringent fiber. In real systems, the degree of birefringence, and the orientation of the birefringent axes, varies from place to place along the fiber. This results in a more complex effect on the optical signal, which is characterized by the concept of “principal states of polarization” or PSPs. PSPs are defined as the two polarization states that experience the maximum relative DGD, and they uniquely characterize the instantaneous state of the system.
Polarization mode dispersion (PMD) is the distortion arising from the statistical sum of the different group velocities of the two components of polarization as the signal propagates through the different sections of the optical communications system. PMD includes first order PMD and higher order PMD and is non-deterministic. First order PMD is the differential polarization group delay at a given wavelength. The instantaneous value for a long fiber can vary over both long time intervals (due to slow variations, such as temperature drift) and short time intervals (due to fast variations, such as mechanical vibration induced polarization fluctuations). The coefficient describing the mean value of first order PMD can vary from >2 ps/km1/2 for relatively poor PMD performance fiber to <0.1 ps/km1/2 for relatively good PMD performance fiber.
Second order PMD arises from two sources: i.) a first order PMD that varies with wavelength; ii.) a change of the system PSP (principal state of polarization) orientation with wavelength, which results in a variation of PMD with wavelength. Second order PMD results in a wavelength dependent group delay, which is equivalent in effect to variable chromatic dispersion, and, can have either a negative or positive sign. The speed of fluctuation is similar to that of first order PMD.
Dispersion imposes serious limitations on transmission bandwidth, especially across long distances, such as in transoceanic routes. Dispersion issues become much more important at higher bit rates, where the separation between the optical pulses is less and where shorter pulses result in a wider signal spectral bandwidth, exacerbating chromatic and second order PMD effects. At bit rates greater than or equal to 40 Gb/s, even for good fiber (<0.1 ps/km1/2 PMD) long length links are deemed to require PMD compensation. PMD can become an inhibiting factor either limiting overall system length or increasing system costs due to the need for additional optical-to-electrical-to-optical signal conversion sites to permit electrical signal regeneration.
One approach to compensation for first order PMD is to introduce a DGD of equal magnitude and opposite sign to the first order PMD in the system. In general, time delays in an optical system can be described in terms of optical path length (OPL) defined byΔ=nL   (2)where Δ is the OPL, L is the physical length of the medium, and n is the index of refraction of the material.
As may be appreciated from equation (2) above, the OPL of an optical waveguide may be lengthened by increasing the index of refraction of the medium or by increasing the physical length of the waveguide. Similarly, the OPL of an optical waveguide may be shortened by decreasing the index of refraction or by decreasing the physical length of the waveguide. Thus, to generate a DGD for PMD compensation, the two orthogonal PSPs of the signal can be sent down two separate paths with different OPLs. If the delayed polarization of the signal is sent down a path with a shorter path length than the leading polarization of the signal, the amount of differential group delay between the two polarizations will be reduced.
A variety of alternatives have been presented to attempt to compensate for first order PMD effects. One proposed system includes a polarization controller and a length of high birefringence polarization maintaining (PM) fiber. A photodetector samples the output signal and attempts to drive the controller using control loop techniques. A long coil of PM fiber (e.g., 50 meters) is necessary to achieve adequately large DGD for dispersion compensation. More importantly, the amount of PMD correction is fixed because of the fixed DGD of the PM fiber, limiting the adaptability and applicability of the system.
Another proposed system attempts to address the problem of adaptability by employing a movable prism element, which generates a variable DGD by varying the distance traveled by one polarization. There are a number of disadvantages to this scheme. For example, optical path losses must be very closely balanced to prevent polarization-dependent loss (PDL). In addition, the overall speed of the variable delay element will be slow due to the mechanical movement of the optics. Furthermore, since the variable delay is created outside of fiber there may be issues of cost and stability due to the complexity associated with the required active alignment of the optical beam.
Another proposed approach to a variable DGD element consists of a single non-linearly chirped grating in a PM fiber. Chirped gratings are gratings in which the spacing of the grating elements varies with position along the grating, so that the effective position at which a signal is reflected depends on its wavelength. In this case, the application of axial strain to the fiber changes the reflection location of each polarization at a different rate, thus changing the delay between them. However, such a design can only achieve a limited range of delays because the differential delay is proportional to the small birefringence of the fiber, and is limited by the small range of strain that the fiber can withstand before breaking. Additionally, this approach induces a varying chromatic dispersion that must be separately compensated.
Yet another proposed approach to generating a differential group delay (DGD) consists of a polarization beam splitter coupled to a pair of single-mode (SM) optical fibers each having a linearly chirped Bragg reflection grating and a controllable extension means for differentially axially straining the fibers. This dual grating approach has the advantage that the two polarization components experience the same chirp when reflected, thereby experiencing matched chromatic dispersion so that polarization-dependent chromatic dispersion is not introduced. However, such a system does not account for polarization fading effects which will require dynamic polarization control in each of the SM fiber grating paths to assure that all light returns properly through the polarization splitter; this will substantially add to the complexity and cost of the system. Furthermore, such a system does not address the difficulty of balancing the two legs during manufacturing, or of properly biasing the system when the compensator is first turned on.
The need remains for a reliable, wide-dynamic range, dynamically tunable PMD system.